Fix GPU MPI coords device pointer bug#186
Merged
awvwgk merged 1 commit intowavefunction91:skalafrom Mar 16, 2026
Merged
Conversation
In the multi-rank MPI path, prepare_onedft_features was called with CPU TensorOptions but passed device_data_ptr->coords_device_data(), which is a GPU pointer. This is undefined behavior — a CPU tensor wrapping device memory. Build host_coords from the Molecule object instead, which is already available on the host and contains the same atomic coordinates. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
In the multi-rank MPI path,
prepare_onedft_featureswas called with CPUTensorOptionsbut passeddevice_data_ptr->coords_device_data(), which is a GPU pointer. This is undefined behavior — a CPU tensor wrapping device memory.This fix builds
host_coordsfrom theMoleculeobject (already on the host) instead of using the device pointer.We probably missed this because the multi-rank case is not as well-covered by tests as the single rank path?